Abstracts

Rationale: Current techniques for the optical investigation of seizures are restricted by light penetration to within 1-2 mm of the cortical surface, an obstacle to evaluation of deep epileptic foci. Previous research has suggested that optical changes, specifically scattering coefficient, may precede electrographic seizure onset in in vivo models. These changes may be related to preictal glial swelling. We describe the development of an implantable fiber-based optical probe to detect preictal optical scattering changes from the in vivo rodent hippocampus. We also report altered optical characteristics during seizure in transgenic mice lacking the water channel aquaporin-4. Methods: Our optical reflectance probe utilized a broadband near-infrared (NIR) illumination source and a NIR-enhanced digital photodiode array spectrometer for signal acquisition. Reflectance values at 800 through 900 nm wavelengths were recorded. The probe was verified using intralipid phantoms with known optical scattering coefficients as determined by frequency domain photon migration spectroscopy. The reflectance probe and a bipolar EEG electrode were stereotactically co-implanted in the hippocampus of adult CD1 mice. Generalized seizures were induced using pentylenetetrazol (PTZ) (100 mg/kg, IP). Simultaneous hippocampal optical and EEG tracings were recorded. Seizures were reversed with pentobarbital (PB) (30 mg/kg, IP). EEG seizure onset was determined post-hoc by an observer blinded to the optical data and separately by power analysis of EEG epochs. Identical experiments were carried out in age-matched transgenic aquaporin-4-null (AQP4-/-) mice and the magnitude and rates of the optical signal change were compared with the wild-type (WT) controls. Results: Reflectance values of the implantable probe were highly correlated with scattering coefficient in the intralipid phantoms (Rsquare = .998) (Figure 1A). All seizures were accompanied by a 2-4% decrease in the scattering dominated reflectance signal (e.g. Figure 1B). Reflectance values recovered following seizure termination. A statistically significant change (>2 S.D. from pre-treatment baseline) in optical reflectance was seen in all mice prior to EEG onset (lead time 82-155s; mean: 123s) (Figure 1C). AQP4-/- mice had lower magnitudes and significantly lower rates of optical signal change compared to WT controls (Figure 2). Conclusions: We have detected optical changes from the hippocampus during the preictal period with a stereotactically implanted fiberoptic probe. Ictal scattering coefficients obtained with these deep optical recordings were of similar magnitude to those obtained in superficial cortex with modulated imaging and previously reported (Owen et al., AES 2007). In this model, optical change reliably precedes electrographic seizure onset by tens of seconds. The slowed kinetics of the optical change observed in AQP4-/- animals supports the hypothesis that glial swelling and reduction of extracellular space may partially underlie the optical signal changes. Further evaluation and optimization of this probe for use as a seizure detector is warranted.